Introduction
Twitter has announced that it is open-sourced its recommendation algorithm. The entire technology and developer community has started developing & improving the algorithm.
The Twitter recommendation algorithm is now live on GitHub. The organization has recently open-sourced some of the most exciting algorithms, including its recommendation algorithm on GitHub. Developers anywhere in the globe can now access and use the system for free.
As expected, the repository was flooded with tons of improvements and bug fixes from the developer community within its short tenure of release. While some developers have fixed many major bugs, some are making fun of the algorithm by trolling with memes and jokes.
The source for the recommendation algorithm can be accessed by everyone, who will simply have a GitHub account. Although Twitter has open-sourced its recommendation algorithm, it doesn’t provide any traces even about the advertisement recommendation algorithm.
The recommendation algorithm is undoubtedly very complicated as it is made up of multiple models of the application. The organization has also released a technical post on its official blog regarding the recommendation pipeline.
According to Twitter, the recommendation algorithm functions nearly 5 billion times every day. It is designed to recommend half of the posts from the individuals we follow and the remaining 50% from unknown organizations and even individuals.
- In November 2022, the Twitter algorithm was changed and designed to recommend tweets more from the people who are not following them.
- This change has almost ruined the entire user experience of the Twitter community, and soon the organization has reverted to the initial recommendation system.
Twitter CEO Elon Musk has said that the code is open-sourced to show the radical transparency of the organization and its recommendation algorithm functions without any malpractices. He named this decision to be another great step towards the encouragement of volunteer developers to access and work on the algorithm to improve its efficiency.
How does the Twitter recommendation algorithm work?
- Twitter has even pointed out how its recommendation algorithms work. It follows mainly three key steps.
- At first, the tweets are fetched from a source and the process is called candidate sourcing.
- Next, a machine learning model works on the scraped data and ranks them based on their relevance.
- Finally, it applies heuristics and filters to remove the already watched, interacted tweets and finally displays them to users.
Conclusion
The recommendation algorithm is out for developers to interact with and apply their styles and taste. Everybody is excited about how the developer community is going to reshape the algorithm and the innovations which will be launched with the recommendation algorithm.
By
